hf-legisqa / app.py
gabrielaltay's picture
pinecone test
23a84f2
raw
history blame
1.34 kB
from langchain_community.vectorstores import Pinecone
from langchain_openai import ChatOpenAI
from pinecone import Pinecone as PineconeClient
import streamlit as st
st.set_page_config(layout="wide", page_title="LegisQA")
CONGRESS_GOV_TYPE_MAP = {
"hconres": "house-concurrent-resolution",
"hjres": "house-joint-resolution",
"hr": "house-bill",
"hres": "house-resolution",
"s": "senate-bill",
"sconres": "senate-concurrent-resolution",
"sjres": "senate-joint-resolution",
"sres": "senate-resolution",
}
OPENAI_CHAT_MODELS = [
"gpt-3.5-turbo-0125",
"gpt-4-0125-preview",
]
def load_pinecone_vectorstore():
model_name = "BAAI/bge-small-en-v1.5"
model_kwargs = {"device": "cpu"}
encode_kwargs = {"normalize_embeddings": True}
emb_fn = HuggingFaceBgeEmbeddings(
model_name=model_name,
model_kwargs=model_kwargs,
encode_kwargs=encode_kwargs,
query_instruction="Represent this question for searching relevant passages: ",
)
pinecone = PineconeClient(api_key=st.secrets["pinecone_api_key"])
vectorstore = Pinecone.from_existing_index(
index_name=st.secrets["pinecone_index_name"],
embedding=emb_fn,
)
return vectorstore
docs = vectorstore.similarity_search_with_score("artificial intelligence")
st.write(docs)